The Principal Investigator (PI) proposes to continue studies to improve autologous and allogeneic bone marrow transplantation (BMT) therapy for chronic myelogenous leukemia (CML). During the current funding period the PI has demonstrated that autologous BMT using marrow treated ex vivo with recombinant human gamma-interferon (rIFN-gamma) provides hematologic remissions as well as complete or partial cytogenetic remissions in the majority of patients. In Specific Aim 1, the PI proposes to study the addition of in vivo anti-leukemia therapy in the post-BMT interval as a method for increasing the proportion and duration of complete cytogenetic remission following autologous BMT. Serial pilot studies are described in which the post-BMT anti-leukemia effects of a) rIFN-alpha; b)rIL2; c)rIL2 + autologous, activated natural killer cells are assessed. The PI has also recently described the anti-leukemia activity of an autologous, IL2-stimulated natural killer cell population termed "adherent lymphokine activated killer" cells or"'A-LAK." In Specific Aim 2, the PI proposes to develop techniques for the efficient recovery of a pure A-LAK population suitable for subsequent autologous post-BMT anti-leukemia therapy. Proposed selection techniques include counterflow elutriation, positive selection by adherence to antibodycoated plastic flasks or immunomagnetic beads and collection of A-LAK progenitors in the post-BMT interval. The PI and his colleagues have also recently described a novel method for selection of benign, primitive hematopoietic progenitors from CML bone marrow. Ibis cell population does express the CD34 antigen observed on hematopoietic stem cells (34+), but does not express antigens associated with committed myeloid or lymphoid lineages (Lin-) and does not express the HLA-DR antigen (DR-). In Specific Aim 3, the PI proposes to develop methods for the efficient collection of benign Lin-34+ DR- progenitors suitable for autologous BMT including counterflow elutriation, immunomagnetic bead separation, fluorescence activated cell sorting, staining with Rhodamine 123 and other physical cell separation techniques. Finally, the PI and his co-investigators have presented preliminary evidence that specific molecular genetic characteristics of the BCR-ABL gene abnormality predict poor outcome following allogeneic BMT therapy. In Specific Aim 4, the PI proposes to continue studies determining the association between pre-BMT origin of BCR-ABL and post-transplant hematologic relapse, determining the association of BCR-ABL mRNA splice patterns by PCR amplification with post-BMT hematologic relapse and correlating the persistence or recurrence of minimal evidence of recurrent leukemia by PCR amplification of mRNA with hematologic relapse.